Approximation-assisted point estimation

Authors:
Barry L. Nelson;Bruce W. Schmeiser;Michael R. Taaffe;Jin Wang
Affiliations:
Department of Industrial Engineering and Management Sciences, Northwestern University, Evanston, IL 60208-3119, USA;School of Industrial Engineering, Purdue University, West Lafayette, IN 47907, USA;Department of Operations and Management Science, Faculty of Scientific Computation, University of Minnesota, Minneapolis, MN 55455, USA;Department of Mathematics and Computer Science, Valdosta State University, Valdosta, GA 31698-0040, USA
Venue:
Operations Research Letters
Year:
1997

Citing 1
Cited 5

Time-dependent queueing network approximations as simulation external control variates

Operations Research Letters

Bayesian analysis for simulation input and output

Proceedings of the 29th conference on Winter simulation
Comparison of Bayesian and frequentist assessments of uncertainty for selecting the best system

Proceedings of the 30th conference on Winter simulation
Bayesian methods: bayesian methods for simulation

Proceedings of the 32nd conference on Winter simulation
Modeling methodology: new approaches for simulation of wafer fabrication: the use of control variates and calibration metrics

Proceedings of the 34th conference on Winter simulation: exploring new frontiers
Bayesian methods for discrete event simulation

WSC '04 Proceedings of the 36th conference on Winter simulation

Quantified Score

Hi-index	0.00

Visualization

Abstract

We investigate three alternatives for combining a deterministic approximation with a stochastic simulation estimator: (1) binary choice, (2) linear combination, and (3) Bayesian analysis. Making a binary choice, based on compatibility of the simulation estimator with the approximation, provides at best a 20% improvement in simulation efficiency. More effective is taking a linear combination of the approximation and the simulation estimator using weights estimated from the simulation data, which provides at best a 50% improvement in simulation efficiency. The Bayesian analysis yields a linear combination with weights that are a function of the simulation data and the prior distribution on the approximation error; the efficiency depends upon the quality of the prior distribution.